C Values in KSVM

I'm not new to data analysis at an excel level, but I am totally new to R. I have been entering a multiple attribute dataset and have found the coeffcients, intercept, predictability, etc. But I find, almost regardless of kernel, but especially in vanilladot, the predictability doesn't change no matter how small or large I make C. I am pretty sure I am missing something because others have found a point of change. Any recommendations on what I might be missing? Thanks.

In most of the data sets where I have looked, the cost parameter does not impact the model fit for linear kernels.